depends: Switch from multilib to platform-specific toolchains #32162
Conversation
|
The following sections might be updated with supplementary metadata relevant to reviewers and maintainers. Code Coverage & BenchmarksFor details see: https://corecheck.dev/bitcoin/bitcoin/pulls/32162. ReviewsSee the guideline for information on the review process.
If your review is incorrectly listed, please react with 👎 to this comment and the bot will ignore it on the next update. ConflictsReviewers, this pull request conflicts with the following ones:
If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first. |
4bf8e8f to
1925dc1
Compare
| @@ -6,11 +6,12 @@ | |||
|
|
|||
| export LC_ALL=C.UTF-8 | |||
|
|
|||
| export HOST=i686-pc-linux-gnu | |||
| export HOST=i686-linux-gnu | |||
There was a problem hiding this comment.
i've always been confused about the -pc- in the architecture tuple-is it directly related to use (or non-use) of multilib?
There was a problem hiding this comment.
It's hard to say.
Referring to x86_64-pc-linux-gnu is still necessary when building depends natively on x86_64, because:
$ uname -m
x86_64
$ ./depends/config.sub $(./depends/config.guess)
x86_64-pc-linux-gnu
However, dropping the -pc- infix helped to with Clang's paths in the OSS-Fuzz environment.
There was a problem hiding this comment.
Hrm, yea, config.sub adds it for x86 archs (but unknown for others) so it's likely canonical somehow.
$ ./config.sub i386-linux-gnu
i386-pc-linux-gnu
$ ./config.sub x86_64-linux-gnu
x86_64-pc-linux-gnu
$ ./config.sub arm-linux-gnueabihf
arm-unknown-linux-gnueabihf
$ ./config.sub riscv64-linux-gnu
riscv64-unknown-linux-gnu
But seems fine to leave it out in the README.md because it doesn't include the -unknown- infixes either.
There was a problem hiding this comment.
Also, please note that depends/README.md lists triplets after specifically mentioning "for cross compilation". In such cases:
bitcoin/depends/hosts/default.mk
Lines 1 to 3 in 4c1906a
-pc- infix.
|
Concept ACK, multilib is more or less specific to x86, it's better if multi-platform is handled in a more architecture-agnostic and consistent way. |
|
Rebased due to a conflict with the merged #30997. |
Guix builds (on x86_64) [untrusted test-only build, possibly unsafe, not for production use] |
|
Rebased due to a conflict with the merged #32086. |
Guix builds (on x86_64) [untrusted test-only build, possibly unsafe, not for production use] |
|
My Guix build: |
|
🐙 This pull request conflicts with the target branch and needs rebase. |
Using the multilib GCC toolchain, as currently documented in
depends/README.md, has several issues, such as:The
g++-multilibpackage conflicts with platform-specific cross-compiler packages. This means it is not possible to cross compile fori686and other platforms using the same set of installed packages.The
g++-multilibpackage is not available forarm64:This PR addresses all the issues mentioned above by switching from multilib to platform-specific toolchains.
Also see #22456.
Here are examples of building for different scenarions:
x86_64orarm64, building with depends natively:x86_64orarm64, cross compiling fori686-pc-linux-gnu:x86_64, cross compiling forarm64:arm64, cross compiling forx86_64: